Official Rules are given as Preformatted text
The registration phase will last for roughly a month, during which players will be registered and materials distributed. All registered players will then enter the Practice Round, consisting of as many 4-player games as necessary. These games will last 10 turns or 2 months, whichever comes first. The Practice Round does not count in the official results, so you can freely explore the rules. A Moderator will provide active guidance. Only those players who complete the Practice Round will be allowed to continue.
This introduction is designed to get everyone up to speed on tournament play, and to weed out players who are unsuitable for the tournament for some reason. It also makes the tournament more fair since all players will have some experience before being thrown into official games. No prizes in the Practice Round; it does not count in the official standings either.
Each human player is allowed to register once for the tournament. Players found to have registered more than once will be dropped from the tournament and banned from all future Warlords PBEM activities.
The tournament will consist of 3 initial rounds, followed by a series of additional rounds and cuts, until only The Greatest survives. Each official round will consist of enough 4-player, 20-turn games to accomodate all surviving players. Each game will have a designated Moderator to provide rules guidance upon request. The exact structure of Rounds and Cuts will be announced at the start of the First Round; it depends on the number and distribution of players.
Games will start at the same time and play concurrently. Players may decline to play in future rounds if they feel they have no chance to survive the round, but everyone is encouraged to complete any game they start - or lose all semblance of Warlords honor. Players who drop out of games in the middle will seriously disrupt the balance of the game and possibly affect the outcome of the tournament. Players are welcome to drop out of the tournament between games, but there will be penalties for those who drop out within a game - you will be considered dishonorable and unreliable and suffer appropriately in upcoming Warlords play-by-email activities (and we have some exciting things planned for after the tournament!).
Because randomly sorting players into games will not necessarily distribute the "best" warlords evenly among sides, it is important (particularly in the first round) to have multiple games before making a "cut", and to reshuffle players among the different colors from game to game.
The 4-player, 20-turn format with 48-hour turnaround was chosen to allow reasonably complete and interesting games to be played within a reasonable amount of time. Games will take about 4 months each. Fewer players would lead to uninteresting diplomacy, and more players would lead to even longer games. Fewer turns leads to incomplete games, and more turns leads to games that take too long. In playtesting this 4/20/48 rule worked very well, using the speedy Tournament Armies and scenarios around 30-50 cities. Smaller scenarios were often either unbalanced, stalemated, or too chesslike; games with more cities were hard to manage because of the 48-hour rule and the intricacy of combat with WarBOT.
For the initial 3 rounds there will be "Four Groups" of players, corresponding to the four flavors of Warlords: PC-Deluxe, PC-Classic, Mac-CD/Deluxe, and Mac-Classic. Players will be assigned to games randomly while minimizing the number of players who fight the same opponent more than once and maximizing the variety of colors each player has a chance to play.
Each player will be involved in at most 2 rounds at the same time. Round B will start when the slowest game in Round A reaches 10 turns or 2 months, whichever comes first. Round C will begin when Round A ends.
When the first cut occurs, a single pool of players will be chosen for Round D, based on the rankings at the end of Round C. (This pool may be divided into "Classic" and "Deluxe" groups with cross-platform play within each group; this would allow Classic players to play one more round before obtaining Deluxe.) Those who miss the cut will be seeded on the Head-to-Head Warlords Ladder by final rank, and they will also become alternates - and may thus get a second chance in the tournament.
The middle rounds will be played using only the Deluxe/CD version of Warlords. Players will now be seeded according to prior performance, with players from different Groups initially interlaced and evenly distributed. If a player drops out of a late-round game, he or she will be replaced by the highest-ranked alternate available from the same initial group. Each alternate will have 48 hours to respond to this "call to arms"; once an alternate is found, he/she has 3 days to complete the lost player's move. Alternates who enter a round immediately after they were cut will play entirely for themselves; alternates entering later will inherit the ranking of the player they replace, and continue by playing for themselves.
Players will be reshuffled into new color pools for the later games, with seeding again taken into consideration, but ensuring that players do not face the same opponents as in Round D. A series of cuts will be made to select the final four.
Players in all four flavors of Warlords will be ranked on a single scale. You will be "pooled" within your group each round based on the side (color) you play. After each round you will be ranked within your color-pool based on how well you did playing your color. The ranking criteria are:
See the Appendix below for the rationale for this choice of criteria.
Because it's impossible to make a completely balanced scenario, it would be unfair to select only the winners of a given game to advance in the tournament. Instead, the tournament will initially compare only those players who play a particular side on a given map, rather than comparing all players. In other words, if 256 people play a scenario, and we want to choose the top 16, then we should rank the 64 players who play each of the four sides, and choose the top four from each side. What matters is to do well with your side, even if you cannot win the game.
Results should be determined as in Tutorial #8. A reminder that cheating or falsifying results will result in ejection from the tournament and lifetime ban on PBEM play. In case of disputes, players should send the game file to the Moderators for arbitration.
Starting with Round B, your cumulative ranking will be determined. It is based on your average rank-within-pool, combined with an "opponent quality" factor which depends on your opponents' average rank-within-pool for the games where you didn't play them. Details will be explained when Round A begins. Thus, you will fight against other players in your game, but your real objective is to play your color better than anyone plays that color in the other games. But you won't know how those other games are going until they finish, so you'd better play your best!
Prizes will be awarded to the top 16 players, and also for the "Best Role-Play" (individual effort) and "Most Entertaining Game History" (group effort by the four players involved) in each of the first three rounds. At the end of the round players can submit reports of clever battle victories for the Warlords Hall of Fame, and those who wish can have their accumulated Hero slayings tallied in the Roster of Warlords Aces. (Players can nominate themselves or each other, and will be judged by a panel of moderators.) Eventually there will not be enough players to maintain the color-pool system. In the final rounds players will send in their preferences for sides, with the higher-ranked player having priority for a side. Eventually there will be four players left, in the final round. This game will be fought to the death, with no 20-turn limit. The Diplomatic History of this game alone will be made "public" on the remailer so that everyone may observe (but not interfere). Players will be allowed twice as many "timeouts" as usual. The winner of this game is the winner of the tournament! The final four will receive prizes and the winner will be declared "The Greatest!" - at least until the next tournament!
In the final game, the razing option will be determined based upon the scenario to be used. The top-ranked player will choose his side first, then the second-ranked player, and so on.
Appendix: The choice of Ranking Criteria:
Here's the fundamental rationale for the city-fraction rule:
Tournament Games will have two types of outcomes. In the first type, one player is clearly winning on Turn 20. The other players either have no cities left, or have so few that they pose no serious threat. This would be the case if one player wound up with at least 50% of the cities. The other type of outcome is where the game reaches Turn 20, but the final conclusion "if the game were finished" is still in doubt. Since the time restrictions in the tournament prevent games from being played to a final conclusion, we need some way to assess and compare players after Turn 20, regardless of the final outcome.
A player's odds of final victory are roughly proportional to the number of cities they happen to have, and are not very strongly correlated with their current rank. Saying someone is "in 1st place" doesn't tell you how likely they are to stay there for the rest of the game, but saying they have 45% of the cities is generally a pretty good clue. A player who's in first place but is barely ahead of the competition with 30% of the cities is not as likely to win as a player who is in 2nd place with 35% or 40% of the cities. A player who's in a tight race and is in 4th place with 20% of the cities still has a very good chance of turning things around, whereas a player in 4th place with a single city is probably doomed.
Since we have to assign rankings fairly to players for all games which are still "inconclusive" on Turn 20, we want to do it with a scoring system that's got as many available rankings as possible, so that the score will most accurately reflect that player's abilities compared to the other players in the game. The scoring system also needs to be simple and obvious from looking at your own turn (so players can know "how they're doing" at all times).
Warlords gives us 5 general options: rank, cities/territory, gold/income, armies/army strengths, and victories/triumphs/bloodshed. Since war is generally about capturing and holding territory, it's preferable to measure territory directly (cities or rank) rather than measuring gold, armies, army strength, or bloodletting. "Rank" is a very coarse measurement which isn't very flexible. City fraction is much more flexible and conveys the same information as rank, so it's generally preferable and more fair.
For instance, if the outcome of the game is still in doubt on Turn 20, players with more *cities* (not necessarily better rank) will generally have a better chance of winning eventually. A player with 42% of the cities on Turn 20 generally has a better chance of achieving a final "victory" than a player with 33%, and it's not fair to the 42% player to give a bonus to the 33% player just because the 33% player is in 1st place while the 42% player is in 2nd place. This is especially true because in close games with a definite cutoff, the last player to move gets an advantage because he can occupy territory without having to worry about retribution. City fractions combined with color-pools can account for this in a way that simple rankings never will. If Kay and Max went into Turn 20 with Max holding a slightly higher city fraction and a stronger position, it wouldn't be fair to give Kay the "1st place finish" just because he was able to snatch 5 cities on his last move which no one else would have a chance to take back. Because rankings can change a lot more dramatically than city fractions can, Blue players in different games will be artificially scattered in the rankings, even though they might all have comparable city fractions. That's not fair. By recording Kay's city fraction and comparing it with other Blue players in other games, that imbalance is automatically factored out. It also avoids the problem where too many Blue players end up with first-place finishes and therefore ties must be broken based on city fractions, while very few Red or Yellow players end up with first-place finishes. Some Yellow player with 30% of the cities could rank "1st" in his pool just because he was the only one to get a 1st-place finish, even though half a dozen Yellows finished with more cities, but in second place behind a Blue player who grabbed a bunch of White and Red cities on his final move. If cities are going to matter a lot for Blue players who tend to finish first, they'd better matter just as much for Red or Yellow players.
Therefore city fraction seems to be the only reasonable scoring system. It measures the right things, it's flexible, it's fair, and everyone is used to it.
Rank, however, can be used as a tiebreaker to give the benefit of the doubt to a player (say Green in game X19) with 25% of the cities, but 2nd place, over another player (say Green in X25) with 25% and 3rd place.
With regard to army strength: the main reason for going with the base strengths is that if you start counting the bonuses you get into a big can of worms, because the exact bonus a unit has depends on who else is in his stack. That means that how you manage to group your units at the end of the game has a strong effect on your final standings, and that's (a) artificial, (b) confusing, and (c) prone to abuse by those who inflate their standings by regrouping units on Turn 21 when no one else is looking. The base strength isn't an ideal measure of an army's power, but it's not bad. Someone with a lot of good stack combinations generally ends the game with more units anyway, just because he doesn't lose as many in battle. Army strength is generally just a tiebreaker anyway.
One possible alternative is to take the number of units of each type and multiply that by the *cost* of the unit, rather than the strength of the unit. The cost reflects the army set designer's opinion of the value of the unit (more or less) and might make a better "effective strength" than either the base strength or the overall modified strength. But cost is not as easy to look up or remember as base strength, and it tends to be a larger number, making the math more painful. The benefit doesn't justify the extra complexity.
Some comments by Dirk and response by Bob:
[ Ranking criteria: ] >I'm sure it's too late for any more suggestions, but I'll toss this in >anyway, for what it's worth... > >Instead of the winning player being determined strictly by the number of >cities he happens to hold at the end of the last player's turn on turn >20, how about awarding one point for each city the player holds at the >*beginning* of *his* turn, *every* turn, and accumulating those points >for the duration of the game. We thought about this earlier (although it might be the only idea that *didn't* get into the Duel2 discussion), and although "integrating" the player's city count might provide a better representation of ability, that's not entirely clear. Part of me really wants to reward the guy that led that three-way alliance that managed to bring down the big guy at the end. And the truth of the matter is that although some empires were great while they lasted, history really only rewards those that are on top at the end. I mean, the Soviet Union was huge, and so was the Third Reich, but I wouldn't consider Hitler the winner of World War II and I wouldn't consider the USSR the winner of the cold war. The other problem is that even if you convinced me that it was more just to reward the long-running leaders over the final winners, it's still a lot more work to compute city scores all the way through the game. I really think that the primary rankings should be so obvious that you don't need to think about them. Finally, we really want to limit the amount of suffering inflicted upon the moderators at the end of the game when they have to double-check all the results. Calculating army strength totals is bad enough! >Advantage: a player who plays well during >most of the game and gets crushed under the weight of three new heroes >during the last few turns will still score higher than a bad player with >the same color in another game who joins in such a three-way carnage and >manages to gain just enough cities to edge him out on that particular >turn. Basically, your Warlord Score would be the area under the graph >of your cities during the game, not just the height of the last dot at >an arbitrary cut-off point. (Should it be "He who dies with the most >toys, wins" or should it be "He who PLAYED with the most toys during his >life, wins"? I favor the second, even if he breaks all his toys in the >last month of his life.) I tend to disagree. If one player is doing well early on and someone else manages to mobilize an alliance to trip him up, I think that shows strong diplomatic skills as well as enough Warlords ability to defeat the big guy. Maybe the guy who does well early on should win in a head-to-head game, but if he can't manage his diplomacy well enough to avoid getting slaughtered by all his enemies in the end, I don't think he deserves to win the tournament. I'd hate to be handing out prizes to players who were great at getting off to quick starts but never actually won a game! >Advantage: it would make the tourney games >more like normal Warlords by eliminating most of the "planning for the >last turn" mental attitude, like "will there be enough time to recover?" >and "I'll just bring my hero back and retake all those cities on the >last turn" and "sack everything, get the gold because that counts". I >think this would give a much more realistic measure of a player's skill >at the game. But normal Warlords doesn't declare you the winner just because you held 40% of the cities for all but the last 3 turns before you got wiped out. It declares you the winner because you got over 50% of the cities on what became the last turn of the game. I agree it would be nice to get rid of some of the end-of-the-game mentality, but I still think in this case the cure is worse than the disease. >It would also be far less likely to cause ties in player >rankings compared to the current system. I wish it wasn't too late to >consider this possibility. (As I said, I haven't read all the other >game logs, so maybe it has already been considered and rejected.) There's no reason why we couldn't make this another option that switches from one round to another, if players were enthused about it and could convince me that it was worth doing. At the highest level all that actually matters is your rank-within-color for each round. How that rank is determined could be flexible from round to round, although again this is something that really ought to wait until after we eliminate the novice players.